The conclusion first: when a foamed part sinks, nine times out of ten it is not about how much foaming agent you added. It is about three things in the mould that were never adjusted for foaming ā mould temperature uniformity, venting, and gate size. On a PP microcellular foamed battery-pack lower housing we ran, two of four cavities showed sink at the rib positions at T0. The customer's first instinct was to cut the foaming agent. We measured nine points across the cavity instead: the gate side and the far end differed by 12Ā°C, and every sink was at the far end. After reworking the water circuit, the same material at the same expansion ratio produced no sink at all.
The core point: solid parts solve shrinkage with packing; foamed parts solve it with temperature and pressure control. A foamed part's apparent shrinkage is smaller than the same material solid ā typically 60%-80% of the solid value ā but the local sink risk is higher, because the skin is only 0.3-0.8mm thick and there is nothing but bubbles underneath to hold it up.
Small overall shrinkage does not mean no local sink
These two things get discussed as if they were one. Overall dimensional change is the linear shrinkage of the whole part after cooling. Local sink is one small area of surface being pulled down by the shrinkage beneath it. In foamed parts, the two behave in opposite directions.
The reason is structural. A foamed part is a skin-core structure: a dense outer skin over a core of bubbles. The core has low density, less heat per unit volume, and lower crystallinity than solid material, so overall dimensional change really is smaller. But wherever wall thickness steps up and material accumulates, the core bubbles keep collapsing after packing ends, the thin skin cannot resist, and a dent gets pulled into the surface.
| Item | Solid injection-moulded | Foamed |
|---|---|---|
| Shrinkage value | Material standard, PP commonly 1.5%-2.5% | 60%-80% of the solid value, must be confirmed by trial |
| Surface skin | Dense throughout | 0.3-0.8mm, weaker resistance to sink |
| Cavity temperature spread | Ā±10Ā°C usually acceptable | Tighten to Ā±5Ā°C |
| Vent depth | Commonly 0.02-0.03mm | One step deeper, commonly 0.04-0.06mm |
| Cooling time | Baseline | Bubbles insulate; estimate 1.3-1.5Ć the solid part |
Troubleshooting order: mould temperature ā venting ā packing
When something goes wrong on the floor, do not start by changing the material. Walk these three steps and you can usually localise it within half an hour.
Step one: measure mould temperature. Take a contact pyrometer to nine points across the cavity ā front, back, left, right and centre. If the spread exceeds 5Ā°C, the sink position will follow the cold side. Where the mould is cold, melt pressure decays fast at the flow end, the bubbles freeze before they can distribute evenly, and shrinkage concentrates there.
On water circuit layout, foamed moulds usually keep cooling channels 10-15mm from the cavity surface with 30-40mm spacing, and deep cavities or rib-dense areas need their own circuit. Our battery-pack lower housing had a single series circuit covering the rib area, so the far end was barely cooled; we added a dedicated in-and-out pair before the sink came under control.
Step two: look at venting. Foaming releases gas, so the amount of gas in the cavity is far greater than in solid moulding. Keep the vent at 0.02mm and the gas cannot escape, creating trapped gas at rib troughs and around boss columns ā visible as local glossy or dull patches, or a whole section collapsing. Foamed mould vents go to 0.04-0.06mm with 3-5mm width, placed at the end of material flow and at the end of every rib trough. Deep vents flash easily, so parting-plane fit accuracy has to rise with them; the two changes belong together.
Step three: reduce packing pressure. This is the counter-intuitive one. Solid parts rely on high packing to compensate shrinkage, and 60MPa or more is normal for PP. Give a foamed part that much and the bubbles are crushed smaller, local density rises, and the part starts behaving like a solid one ā shrinkage goes up and sink follows. Foamed parts usually take half the solid packing pressure or less, and the higher the expansion ratio, the lighter the packing.
On gates: point gate diameter is enlarged 20%-30% versus the same solid part ā a small solid part running 1.0-1.2mm often starts from 1.5mm when foamed. Too small a gate means shear heating makes the material start foaming near the gate, so it is already unevenly distributed by the time it reaches the cavity.
What has to change in the drawing and the mould
A foamed part's drawing cannot be drawn from solid-part habits. These are the places we have revised repeatedly.
Rib thickness at 0.6-0.7Ć the main wall. Solid parts are usually pushed to 0.5-0.6Ć; foamed parts can relax slightly because the low-density core packs less material at the same thickness. The rib root must carry a radius of at least 0.5mm ā it spreads stress and gives material a flow path.
Draft angle increased 0.5-1Ā°. A foamed surface has micropores, so friction against the mould is higher than a solid part; too little draft scuffs the surface or tears the skin. Deep parts we usually take to 1.5Ā° or more.
Keep ejector pins off thin-skin areas. Foamed skin has low strength, so pins on a large flat face will whiten or punch through. Put them on ribs, bosses and flanges where there is structure, and increase pin diameter a little.
Measure trial dimensions 24 hours after ejection. After ejection, gas slowly escapes from the core bubbles and dimensions keep moving; readings taken in the first few hours mean nothing. Shrinkage must be set from dimensions taken after 24 hours or longer, otherwise the mould you cut will produce yet another size after a week in the customer's warehouse.
Root-cause criteria: if the sink is in the same place every shot, it is a structural issue ā change wall thickness and radii. If it drifts from shot to shot and comes and goes, it is mould temperature or the water circuit. If the same cavities fail every time in a four-cavity mould, it is runner balance. If the whole batch is uniformly undersized, the shrinkage value is wrong or post-shrinkage was not accounted for.
Questions we get most often
Q: What shrinkage should be used for a foamed part?
A: Usually 60%-80% of the shrinkage for the same material in solid form. PP solid is commonly 1.5%-2.5%, which puts a foamed part roughly in the 1%-1.8% range, and the trial data decides. Higher expansion ratio gives lower shrinkage, but the relationship is not linear ā mould temperature, packing and wall thickness all matter. Run at least three expansion ratios from the same mould, measure after 24 hours, and only then fix the shrinkage for cutting steel.
Q: Where does sink usually appear on foamed parts?
A: Wall-thickness steps, rib roots, boss bases, the far end opposite the gate, and trapped-gas dead corners such as rib troughs. What they share is a lot of material underneath and a thin skin. The approach is to remove accumulated material first ā bring rib thickness toward 0.6-0.7Ć the main wall, add radii at rib roots, then check whether the vent path at those positions is open. Only when the structure cannot be changed should you move to mould temperature and packing.
Q: Three steps feel like a lot. Which one first?
A: Mould temperature. It is the fastest to measure and the most common cause. Nine points with a contact pyrometer takes minutes, and if the spread exceeds 5Ā°C you already have your answer in most cases. Venting comes second because it explains the localised glossy or dull patches that temperature does not account for. Packing pressure comes last because changing it affects density and mechanical behaviour, and it is the easiest to overshoot.
Q: We added foaming agent and it still sinks ā is the material the problem?
A: Usually not. Measure the cavity temperature distribution first, then check vent depth and the packing setting; rule those three out before going back to formulation and foaming agent ratio. Packing is the most overlooked of the three: give a foamed part high packing and the bubbles are crushed, local density approaches that of a solid part, and the shrinkage behaviour follows ā everyone is used to solid parts relying on high packing, so it is easy to go the wrong way here.
Q: How deep should venting be on a foamed mould?
A: Commonly 0.04-0.06mm, one step deeper than the 0.02-0.03mm typical of standard injection moulds, with 3-5mm width, placed at the end of flow and at the ends of rib troughs. Deeper vents flash, so parting-plane fit accuracy must be raised at the same time ā the two changes go together. Vents also need periodic cleaning; foamed material leaves residue in the slots, and once they clog the sink comes back.
Originally published by the Hezi Industrial Design team, based on our industrial design and structural engineering practice in Dongguan. Learn more: Industrial design services
